4. Discussion
This is the first study to describe in detail the absolute phylloquinone intake by preterm infants from all sources during their first weeks of postnatal life. We quantified the relative proportional intake of phylloquinone from the various nutritional routes for infants who received PN and for those who did not receive PN. We found no differences in overall intakes between the three study groups that had been randomly allocated different phylloquinone regimens at birth. However after correcting for birth weight and days to study completion, the average daily intake of phylloquinone adjusted for birth weight was significantly lower for infants who had received the lower dose prophylactic bolus dose of 0.2 mg phylloquinone. We had speculated that the initial bolus doses of phylloquinone given for prophylaxis at birth would assume less importance as a source of the vitamin contributing to the overall intake at study completion, due to the increasing respective contributions to overall phylloquinone intake derived from other nutritional (dietary) sources. In infants who did not receive any PN, the bolus dose given at birth represented approximately 50% of the overall phylloquinone intake at study completion, compared to only approximately 20% in infants who had received PN.
Preterm infants represent a neglected group with respect to current knowledge of their phylloquinone intakes and recommendations for their nutritional requirements. As pointed out by Kumar
et al. [
14], published recommendations for vitamin K intakes and/or supplementation are specific for all ages except for preterm infants. As a result, recommendations for preterm infants are arbitrary and historically have ranged from 5 to 10 μg/kg/day in 1993 [
15] and as high as 100 μg/kg/day in 1988 [
16]. In this context it is important to note that the principles and knowledge base of dietary recommendations have changed over time. For example, the recommendations of 5–10 μg/kg/day for preterm infants in 1993 [
15] came out during the lifespan of the 10th Edition of the US Food and Nutrition Board guidelines published in 1989, at which time they took the form of Recommended Daily Allowances (RDA) instead of the current Dietary Reference Intakes (DRI). At that time the RDA for term infants over the first six months was 5 μg/day, which in turn was based on the adult RDA of 1 μg/kg/day [
17]. This means that the 5–10 μg/kg/day recommendation for preterm infants by Greer
et al. [
15] was 5–10 fold greater than the intake recommendations for term infants at that time. More recent guidelines published in 2005 recommended a phylloquinone intake of 8–10 μg/kg/day as an AI for preterm infants [
18], representing the top end of the 1993 recommendations [
15].
There is an inevitable degree of arbitrariness of recommendations even for healthy term infants. In the current United States recommendations published in 2001, the AI for healthy term infants is based on an average daily milk intake of 0.78 L and an average phylloquinone concentration in human milk of 2.5 μg/L [
7]. This gave an AI of 2.0 μg/day after rounding [
7]. The weak link in this calculation as a precise AI value lies with the fairly wide variations in the reported concentrations of phylloquinone in breast milk which taking the lower and upper values would result in estimated phylloquinone intakes of ~0.5 μg/day and 2.5 μg/day respectively [
8].
Another weakness of current recommendations is that they are based solely on the phylloquinone content of breast milk. Although phylloquinone is the major vitamer of vitamin K in breast milk, it also contains a member of the menaquinone series, namely menaquinone-4 (MK-4) at concentrations that are about half that of phylloquinone [
19,
20]. There is also evidence that this MK-4 in breast milk is derived from dietary phylloquinone [
20]. Future neonatal AI recommendations should also consider the contribution to intakes made by MK-4 as well as its potential contribution to neonatal vitamin K status.
Whether the dietary guidelines have been in the form of an RDA or AI, a common underlying assumption is that the infant is also given a prophylactic dose at birth in amounts recommended by the relevant paediatric societies [
7,
17]. For any individual infant, the optimal daily ongoing phylloquinone required from feeding will depend in part upon what prophylactic dose was given at birth and any postnatal supplementation received from PN. However there is also likely to be inter-individual variation in storage and metabolism.
For the purposes of discussion we have taken the most recent recommendation of 8–10 μg/kg/day [
18] as a benchmark against which the phylloquinone intakes in this study can be compared. The median phylloquinone intake for infants who completed this study was approximately three-fold this recommended amount, and up to five to seven-fold more in some infants (
Table 3). For infants who received PN, there was little difference between allocation groups. The median of the average phylloquinone intake was three to four times the recommended amount, and all had an average phylloquinone intake that ranged between ~20–70 μg/kg/day,
i.e., two to seven times the daily recommendation (
Table 4). For infants who did not receive any PN, the median average daily intake received by 0.5 mg IM group infants was approximately twice that currently recommended, whereas for infants in the 0.2 mg IM and 0.2 mg IV bolus groups the medians and ranges of average daily intakes were remarkably close to the currently recommended intakes of 8–10 μg/kg/day (
Table 5).
For infants who did not receive any PN, the bolus dose(s) of phylloquinone constituted the major source of phylloquinone intake in the study period. In contrast, for infants who had received a period of PN by far their major intake source was, somewhat surprisingly, that delivered by the PN solution. For comparison, the ratio of median phylloquinone intake from PN to that from bolus doses was ~3:1.
In infants given PN, the median overall phylloquinone intake from enteral feeds was 6% (range: 0%–28%) of the total intake. In contrast in the infants not given PN the proportional intake from enteral feeds was significantly higher, and comprised 23% of overall intake for infants receiving a 0.5 mg bolus dose of phylloquinone after delivery, and 50% of overall intake for infants who received a 0.2 mg dose. Thus the contribution towards overall phylloquinone intake made by enteral feeding is small, but becomes more important in infants not given PN and particularly so for infants who receive lower bolus prophylactic doses (0.2 mg) and who do not receive PN.
These data show that in almost all infants, the average daily phylloquinone intake was in excess of the currently recommended amounts. The intakes were particularly high at three to four times the recommended amounts in all who received PN (irrespective of prophylactic dose), and were approximately twice the recommended amounts in infants given 0.5 mg prophylactic doses who did not receive PN. Only infants who received the 0.2 mg dose and who did not receive PN had an intake which approximated well to the current recommended daily phylloquinone intake of 8–10 μg/kg/day. Our data provide further evidence that the amounts of phylloquinone currently added to manufactured PN multivitamin solutions are excessive for the needs of preterm infants and should be reviewed.
Although there are no known clinical manifestations of toxicity from phylloquinone prophylaxis in neonates, nor indeed any known toxic effects in adults consuming high amounts of vitamin K [
7], we previously found evidence of hepatic overload in a sub-group of preterm infants from our main study who had received the 0.5 mg rather than the 0.2 mg dose for prophylaxis [
9]. The larger dose was more likely to be associated with elevated serum phylloquinone 2,3-epoxide concentrations, and suggested overload of the hepatic vitamin K 2,3-epoxide reductase enzyme in some preterm infants [
2,
9]. Furthermore, in a separate study we previously obtained evidence of metabolic overload in a subgroup of preterm infants who predominately excreted the less extensively metabolised urinary 7C-side chain metabolite of vitamin K instead of the usual 5C-side chain metabolite [
21]. This subgroup also had the highest serum concentrations of phylloquinone together with raised blood concentrations of phylloquinone 2,3-epoxide that is normally undetectable. This combination of increased excretion of the 7C-metabolite, high serum phylloquinone and raised phylloquinone 2,3-epoxide is indicative of a metabolic overload of both vitamin K recycling and catabolic pathways [
21]. None of the infants in these studies who had biochemical evidence of overload showed any clinical manifestations of an overload status.
Because of their special nutritional requirements preterm babies are often provided with multivitamin supplements—including the fat-soluble vitamins A, D, and E—at the time of discharge home from the neonatal unit; these are usually continued throughout infancy. However post-discharge phylloquinone supplements are rarely provided to this group at present. Yet without adequate ongoing supplementation preterm infants may be at increased risk of developing vitamin K deficiency in early infancy, particularly if they continue to be exclusively breast fed. National surveillance studies of VKDB continue to report sporadic cases in preterm infants [
6]. The most recent study in the United Kingdom and the Irish Republic reported a case of probable VKDB in a 24-week gestation infant who had received 0.4 mg/kg IM at birth [
22]. This infant was primarily human milk fed and had no liver disease but suffered gastro-intestinal bleeding aged three months postnatal. Further study is required to assess whether phylloquinone supplements should be given routinely to preterm infants on discharge home from the neonatal unit.